Jamy Martin
I am pursing MS degree in Data Science at École Centrale de Lille.
I am currently an AI/DS Intern at Skim AI Technologies with a focus on NLP and Deep Learning research.
April 2026 • 5 min read
This project builds an end-to-end ML pipeline that forecasts French electricity consumption 48 hours ahead. It automatically ingests daily data from the RTE API, stores 55,000+ hours of historical records in PostgreSQL, and generates predictions using a fine-tuned Chronos-2 foundation model, achieving a 3.04% MAPE on unseen data.